On Fri, 2005-07-01 at 20:08, Richard Kelsch wrote:
> My only beef with Fedora was why was it
> (gcc4) implemented without sufficient warning of its possible effects.
> In fact, the release notes just briefly mention a difference in using
> "lvalues" in an illegal way and that's it. Ok, what's than mean for me
> as a user?
It means that when you try to install code from other places that
uses that illegal technique, it won't compile until someone fixes
it. Which may not be until the day after the author installs a
distribution that includes gcc4.
> Are there any places to go to perhaps to learn how to correct any
> problems resulting from the gcc upgrade?
Sending a bug report to the author of the code that doesn't compile
might help.
> I understand that the Fedora
> team most likely did the change for performance reasons, and not to
> throw a monkey wrench into the community just because they could.
> Please let this be clear.
>
> Constructive enough for you?
No, the Fedora team just included the current upstream package. That's
the question: if upstream packages change in a way that isn't
ever going to be completely backwards compatible, should a
distribution with a fast-release philosophy push the change out
quickly and let everyone get started with the inevitable adjustments
or should they hold back and prolong the pain? I've come down on
the 'hold back' side about some other things, but those were things
that might have been fixed by waiting.
--
Les Mikesell
lesmikesell@xxxxxxxxx
